Weirdly there has been a sudden surge of hiring in between the ai tech giants, a hunt for philosophers with PhDs. Why suddenly? These two factors never merged before in fact they are the literary contrast of a passionate artist and the coder who started at 10. Yet suddenly in this time and age the most advanced industries in entire mankind’s history are standing at the doorsteps of philosophy department for answers. Or should I say questions on the implication of what will be the line in the ai bubble and by what time it bursts. Turns out string of 0s and 1s cannot teach the machine after all to how to think with conscience.
The reason is not that hidden in file stacks, but it is plain and evident in day light ai have been integrated to each of our lives too much that I can’t even think of making another dish for supper without talking to ChatGPT. it's everywhere and mind I call it a plague or a modern-day invention that’s going to turn around? And yet HR officials are not hiring anymore instead its ats running Ai, personal assistants are expensive with 28k per year with an unpaid college degree debt, so AI have a chance to be the cheaper alternative and indeed my surgeon is also a robot.
So, question is a large language model without conscience running on internet gained facts is it less efficient or a tad bit more harmful before we sign off each AI with I agree cookies.
And thus, ethicists joined anthropic, DeepMind decided to hire philosophy team and open ai suddenly gave more openings for humanities graduates. So somewhere around the corner the private sector running on its reputation at risk and a sudden surging alarm got to know that philosophical problems before any of us did.
And yet why the public sector and rulemaking bodies haven’t gotten to the same conclusion?
Are the right people in the room?
Running an eye on how AI governance is being done perhaps first one to point out is the EU AI Act. A risk classifying system built on technical threshold capability this may be the most comprehensive regulatory framework to be attempted so far and meanwhile the UK’s Ai safety institute focuses on evaluating capacity of frontier models. The us also uses the US executive order on AI to show agencies to assess national security implications and competitiveness in the economy. These are not just questions at random. Mostly asked by the engineers and economists instead of philosophers though.
The philosopher may mostly ask what should most system be allowed to dictate? Whose value is the model projecting? Who will be held responsible for the damages to be done. And in entirety who will be held responsible for the algorithm’s right or wrong decisions determining If a human gets a job, loan or prison sentence.
While Somone is supposed to ask about conscience, the rule writers are looking in capability and revenue and thus the questions are largely non-existent in a formal government framework setting.
It’s not just the theoretical harm
While the harm is already documented and consistently going on this cannot be considered as a future risk potential.
Apparently, amazon scrapped its own recruitment AI in 2018 after understanding that it rejected CVS with the word “women’s” showing another example of algorithmic hiring tools have been giving disadvantage to candidates in women and minority sections. Racial bias haven been shown to encode and amplify according to the predictive policing systems in the United States. When in the United Kingdom the algorithm to allocate A level grades downplayed grades of student coming from lower income school background, they had to abandon the system in 2020 due to public opposing. Same way the assessments systems for benefits across different countries are done by automated systems making decision which is to sanction for vulnerable people with no to minimal human oversight or right to appeal.
These are not random cases, but this is what happens when system is built on data which mirrors injustice historically and are opened for usage without proper grounding in philosophical department, ethical overview or frameworks in governance which is embedded in values than efficiency.
The paradox of corporate
The core contradiction happening at current momentum is the private sector running purely on profit driven ideologies with more unaccountable and faster passed than how the regulatory framework can keep pace with, yet they are more philosophically literate about AI meanwhile government is not.
Moreover, the inversion is dangerous as when the philosophical work of deciding how ai must behave to the user is done by corporate the interests Favor them and their private gains rather than in the interest of the public. There was a responsible move from anthropic side its responsible scaling policy but after all its made by anthropic for anthropic. The industry lobbying shaped a major part of EU’s AI act. The people with more at stake is how things are answered, communities marginalised, citizen in ordinary capacity, workers whose jobs are on the line by algorithms. Let’s just say even after graduation ats is running my life.
Governance cannot be outsourced to the governed.
How will the philosophically grounded ai governance looks like
The end goal is to never replace a bunch of engineers and economists with philosophers, but it is to make sure the framework for governance are getting the building blocks on values first and questions with technical patterns second. This leads to three things.
Primarily to show mandatory representation of ethics and philosophy expertise on Regulatory bodies for AI. Not just in the format of advising or consulting but with equal participation in decision making with technical experts.
Secondly governance framework that begins with questions, what judgments should be contributed by humans and what by Artificial intelligence before questioning the implementing and capability. For example here the EU AI acts risk tiers help but they mostly classify the technical risks than the ethical ones.
The third and meaningful part being the question of how AI should behave in a world but this is not a technical question with proper answers found by the experts. This is more of a philosophical and yet political question open to everyone. And each of the citizens need a well-informed way and structure was to participate in the answering of I, not just the consultation that comes later after the decision have already been derived.
The machines are already built
Both the machines and the value have already been made. So now its up to someone to decide how they should behave in the world.
That decisions are being made in agencies for regulation, standard bodies, ethics team for large corporates and even government ministries. Mostly by people who are being taught to look forward to what AI can do and not what it should be doing. The philosophical gap in the governance Is not a future problem but rather a present crisis playing out in different formats today about systems already deploying at larger scales.
Philosophers did not build this but definitely have a role in saving it.
"This piece was written with structural assistance from AI. The ideas, argument, and voice are my own."
